Synthesis of DACM, Asp-DACM and Asp-AMS

The synthesis of DACM has been described previously [24,25]. The synthesis of 5′-O-(N-L-aspartate)-sulfamoyl-adenosine (Asp-AMS; Scheme 2 in S1 Text) was performed as detailed in S1 Text. 2-chloro- 5′-O-[N-(L-aspartyl)-sulfamoyl] adenosine (Asp-DACM) was prepared by adapting literature methods [26,27]. Briefly, the bis-protected N-(Boc)-tert-butyl aspartic acid was converted to the N-hydroxysuccinate ester (1) with N-hydroxysuccinimide and N,N’-dicyclohexylcarbodiimide (DCC). Next, the aspartate N-hydroxysuccinimide ester (1) was coupled directly with the 5′-O-sulfamoylated isopropylidene protected adenosine (2) in the presence of 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) in N,N-dimethylformamide (DMF) to yield the triply protected N-aminoacylated sulfamoyl adenosine derivative (3). Lastly, complete global deprotection of 3 with trifluoroacetic acid (TFA) in water and tetrahydrofuran (THF) gave the desired 2-chloro-5′-O-[N-(l-aspartyl)-sulfamoyl] adenosine derivative, Asp-DACM, in high purity after precipitation from acetonitrile (MeCN) and triethylamine (Et₃N) (Scheme 1 in S1 Text).

Effect of DACM on P. falciparum growth, stress response and protein translation, and toxicity to mammalian cells

Synchronized ring stage parasites (3D7 strain [28]) were exposed to increasing concentrations of DACM, and parasite viability was assessed in the next cycle by flow cytometry [29]. DACM is a potent inhibitor of the growth of P. falciparum cultures (Fig 1B; IC_{50_72h} = 3.3 ± 0.1 nM (n = 5)) with an efficacy similar to the previously reported PfTyrRS inhibitor ML471 [16], and to the current front-line drug, dihydroxyartemisinin (DHA) [17]. It should be noted that DACM also exhibits significant toxicity against the mammalian cell line, HepG2, with an IC_{50_48h} value of 47 ± 10 nM (n = 4) (Fig 1B). Adenosine 5′-sulfamate (AMS, S1A Fig), a structurally related compound [15], also exhibits potent activity against P. falciparum cultures (S1B Fig; IC_{50_72h} = 3.8 ± 0.2 (n = 5)). AMS inhibits the growth of HepG2 with a IC_{50_48h} value of 53 ± 2 nM (n = 5) (S1B Fig).

Accumulation of uncharged tRNAs triggers the amino acid starvation stress response. Using established methods [17], we showed that eukaryotic initiation factor 2α (eIF2α) is phosphorylated upon exposure of P. falciparum cultures to DACM and AMS (Fig 1C, left panel). A similar response was observed for ML901, a known reaction hijacking inhibitor of PfTyrRS, as well as the conventional tRNA synthetase inhibitor, borrelidin (Fig 1C, left panel). By contrast, in transfectants in which the eukaryotic translation initiation factor 2-alpha kinase 1 (eIK1) has been deleted [30], DACM and other aaRS inhibitors did not cause eIF2α phosphorylation (Fig 1C, right panel). As previously reported [31], exposure of the eIK1 knockout to dihydroxyartemisinin (DHA) still induces eIF2α phosphorylation (Fig 1C, right panel).

The aaRSs charge tRNAs with amino acids to drive protein synthesis. We monitored protein translation in P. falciparum by following the incorporation of an alkyne analogue of puromycin (OPP), followed by attachment of a clickable fluorophore [32,33]. Upon treatment of trophozoite stage parasites with DACM or AMS, protein translation was inhibited with IC_{50_3h} values of 32 ± 1 and 27 ± 6 nM, respectively (Fig 2A and 2B). These values correlate well with the IC_{50_3h} values observed for parasite killing of 22.6 ± 0.2 and 18 ± 4 nM, respectively (Fig 2A and 2B). These data are consistent with inhibition of protein translation being directly linked to parasite killing. An equivalent 4-h exposure to the folate pathway inhibitor, WR99210, kills parasites without affecting protein translation in the period monitored (Fig 2C), while a ribosome-directed protein translation inhibitor, cycloheximide, inhibits translation, but this short exposure results in limited killing (Fig 2D).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. DACM treatment inhibits protein translation in P. falciparum cultures.

Late-stage P. falciparum (Cam3.II^rev) infected RBCs were exposed to the inhibitors for 6 h with the incorporation of OPP during the final 2 h. For parasite viability, compounds were washed out after the 6-h exposure, and viability was assessed in the next cycle. (A) DACM, (B) AMS, (C) WR99210, (D) Cycloheximide.

https://doi.org/10.1371/journal.ppat.1013057.g002

DACM targets multiple aaRSs in P. falciparum, including PfAspRS but not PfTyrRS

Reaction hijacking nucleoside sulfamates attack the activated oxy-ester bonds of the enzyme-bound aminoacyl tRNA to form tight-binding adducts (Fig 3A). We used targeted mass spectrometry to search for potential conjugates in P. falciparum-infected red blood cells (RBCs) that had been treated with DACM (10 µM, for 3 h). Following Folch extraction of lysates, the aqueous phase was subjected to LCMS and the anticipated masses for the 20 amino acid conjugates were interrogated. Signals were observed with precursor m/z values within 5 ppm of theoretical values calculated for DACM adducts of Asn (most abundant), as well as Asp, Thr, Ser and Lys, while weaker signals were observed for His and Phe (Figs 3B,3C and S2). The identity of the Asp-DACM adduct was confirmed using a synthetic standard, which exhibited essentially the same retention time, precursor ion m/z value and MS/MS fragmentation spectrum as the species generated by P. falciparum (Fig 3B-D). No corresponding peaks were detected in control samples not exposed to DACM. The data indicate that at least PfAsnRS, PfAspRS, PfThrRS, PfSerRS, PfLysRS, PfHisRS and PfPheRS are susceptible to reaction hijacking by DACM.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Targeted mass spectrometry identifies Asp-DACM conjugates in P. falciparum.

(A) aaRSs catalyze nucleoside sulfamate attack on activated amino acids (aa) to form an amino acid adduct. (B-D). Late trophozoite stage P. falciparum 3D7 cultures were exposed to 10 µM DACM for 3 h. Parasite extracts were subjected to LC-MS/MS to search for DACM-amino acid conjugates. (B) Extracted ion chromatogram (XIC) of Asp-DACM (m/z = 496.0648) from P. falciparum extracts and synthetic Asp-DACM standard spiked into untreated parasite lysate. Inset: Structure of Asp-DACM. (C,D) Mass spectra (C) and MS/MS fragmentation spectra (D) of Asp-DACM produced by P. falciparum extracts (top panels) and synthetic Asp-DACM standard (bottom panels).

https://doi.org/10.1371/journal.ppat.1013057.g003

PfTyrRS is less susceptible to hijacking by DACM than by AMS

The lack of Tyr-DACM adduct signal suggested that PfTyrRS is not a target of DACM. This contrasts with AMS, for which Tyr-AMS was the most easily detected product in mass spectrometry data [15]. We generated recombinant PfTyrRS as described previously [17]. We used differential scanning fluorimetry (DSF) [17] to monitor changes in the thermal stability of recombinant PfTyrRS upon incubation with AMS and DACM. Under the conditions examined, AMS exposure caused marked stabilization of PfTyrRS (a 10°C shift), while DACM did not (S3A Fig and S1 Table), suggesting weaker hijacking activity. To understand why PfTyrRS might be less susceptible to hijacking by DACM, we carried out protein–ligand docking using the Surflex fragment matching strategy [34]. Unconstrained docking always resulted in unrealistic orientations for Tyr-DACM, so the position of the Tyr fragment of Tyr-AMP-bound PfTyrRS (PDB: 7ROR) was used to constrain the position of the Tyr-moiety in Tyr-AMP, Tyr-AMS and Tyr-DACM, which were docked into the active site in silico (without protein flexibility). Tyr-DACM gives the least favorable docking score (S2 Table), due to a clash of the Cl atom with surrounding residues (S3B Fig).

Generation of recombinant AspRSs

PfAspRS was chosen for further analysis as it has not been studied as a reaction-hijacking target previously. Alignment of the AspRS amino acid sequences from P. falciparum, P. vivax, Saccharomyces cerevisiae and human reveals the characteristic anticodon-binding domain linked by a hinge region to the C-terminal catalytic domain, with conserved motifs 1 – 3 [35] (Figs 4A and S4). The Plasmodium sequences have a species-specific insertion in the anticodon-binding domain [36], while the Plasmodium and yeast sequences have a variable length N-terminal extension with a lysine-rich motif that is thought to bind RNA [36,37]. Previous studies indicated that translation of PfAspRS is initiated from an internal methionine, Met 49 [36].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Thermal stabilization of native length AspRS enzymes by DACM and AMS.

(A) Schematic diagram of AspRS enzymes from P. falciparum, P. vivax and Homo sapiens. Native length PfAspRS and PvAspRS are initiated from internal methionines, M49 and M51, respectively. The anticodon-binding domain (ABD), the hinge region, Motifs I-III, and the flipping loop are indicated. Truncated PfAspRS and PvAspRS constructs comprise residues 87-626 and 96-631, respectively. (B-G) First derivatives of melting curves for native length PfAspRS (B,E), PvAspRS (C,F) and HsAspRS (D,G) (1.5 μM) after incubation at 37°C for 3 h with 10 μM ATP, 20 μM Asp, 80 μM EctRNA, with 10 or 50 μM DACM (B-D) or 10 or 50 μM AMS (E-G). Data are representative of three independent experiments.

https://doi.org/10.1371/journal.ppat.1013057.g004

We used an E. coli expression system to generate recombinant native length PfAspRS (49-626) (PF3D7_0102900) and the equivalent predicted native length PvAspRS (51-631) (PVX_081610). Full-length HsAspRS (1–501) (NP_001340.2) was also generated to enable a comparison of the susceptibility of the Plasmodium and human AspRS enzymes to reaction hijacking. Following removal of the His-tags, gel filtration yielded dimeric proteins. Given a previous report of poor stability of native length PfAspRS [36], we characterized the preparation by mass spectrometry and analytical ultracentrifugation, showing that PfAspRS exists as a homogenous dimer in solution (S5A and S5B Fig).

We assessed the ability of the recombinant aaRSs to consume ATP in the initial phase of the aminoacylation reaction, using the Kinase GLO assay, as previously described [17]. In initial studies, we found that laboratory prepared E. coli tRNA (EctRNA) is effective as a substrate for all three enzyme preparations, which facilitated the comparison. In the absence of tRNA, very little ATP is consumed by each of the enzymes (S5C Fig). Addition of EctRNA substantively increased the level of ATP consumption (S5C Fig), consistent with productive aminoacylation.

Recombinant P. falciparum and P. vivax tRNA synthetases are thermally stabilized upon formation of Asp-DACM and Asp-AMS adducts

Upon incubation in the presence of substrates, i.e., Asp, ATP and EctRNA, recombinant native length PfAspRS, PvAspRS and HsAspRS exhibited melting temperature (T_m) values of 43.0°C, 43.4°C, and 49.1°C, respectively (Fig 4B–G and S1 Table). These T_m values are 2–3°C lower than for the respective apo AspRSs (S1 Table), which may be due to binding of the SYPRO Orange dye by the tRNA preparation, as reported previously [38].

When PfAspRS was incubated with substrates in the presence of 50 µM DACM, the T_m value increased by 18.1°C (to 61.1°C), consistent with the formation of a very tightly bound Asp-DACM adduct (Fig 4B and S1 Table). At a lower DACM concentration (10 µM), two peaks were evident indicating incomplete conversion to the stabilized form of PfAspRS. PvAspRS appears to be even more efficient at forming the Asp-DACM adduct, with the T_m value shift of 19.0°C (to 62.4°C) already largely complete with 10 µM DACM (Fig 4C). By contrast, very little stabilization of HsAspRS was observed at 10 µM DACM, but an emerging population of adduct-bound HsAspRS was observed at 50 µM (Fig 4D). As expected, incubation with synthetic Asp-DACM increased the T_m values of PfAspRS, PvAspRS and HsAspRS to a similar extent (S5D–F Fig and S1 Table).

Incubation of the recombinant PfAspRS and PvAspRS enzymes with AMS plus substrates also led to marked stabilization, with increases in T_m values to 57.0°C and 59.5°C, respectively. By contrast, HsAspRS was not stabilized by AMS, under the conditions of this experiment (Fig 4E–G and S1 Table). Taken together, these data suggest that HsAspRS is less susceptible to reaction hijacking than PfAspRS and PvAspRS.

Structure of apo PvAspRS

Attempts to crystallize native length PfAspRS and PvAspRS resulted in poorly diffracting crystals. An AlphaFold analysis predicted that the N-terminal extension is likely to be flexible (S6A Fig), which may impede crystallization. We therefore generated truncated PvAspRS (96–631), lacking the N-terminal extension. PvAspRS (96–631) appears to be capable of catalyzing productive aminoacylation, as indicated by a marked increase in ATP consumption in the presence of EctRNA (S6B Fig). PvAspRS (96–631) also appears capable of generating the Asp-DACM adduct when incubated with 50 µM DACM in the presence of ATP, Asp and EctRNA, as indicated by an 18.4°C increase in the T_m value (S6C Fig and S1 Table). Incubation with the synthetic Asp-DACM adduct led to a 20.1°C increase in the T_m value (S6D Fig and S1 Table). These T_m values are similar to those for native length PvAspRS, indicating that the core PvAspRS construct binds the Asp-DACM adduct with similar affinity.

A Morpheus II screen (Molecular Dimensions) was used to generate crystals of PvAspRS (96–631), both as the apo protein and in complex with Asp-DACM, Asp-AMP and Asp-AMS. We solved the structure of apo PvAspRS and refined it to 2.1 Å resolution. Data collection and refinement statistics are summarized in S3 Table. The structure of the apo dimer is presented in Fig 5A, with features highlighted in Chain B, revealing a typical Type II aaRS with an N-terminal β-barrel anticodon-binding domain connected via a hinge to a larger C-terminal catalytic domain that adopts an α-β fold. Motif I is involved in the dimer interface. Motif II and motif III are integral components of the catalytic pocket.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Comparison of the crystal structures of apo and Asp-DACM-bound PvAspRS reveals binding site interactions.

(A) Ribbon diagram of apo PvAspRS (96-631) with Chain A in salmon and Chain B with domains indicated. The anticodon-binding domain (ABD, red), hinge region (yellow), motif I (violet), motif II (green), motif III (cyan) and flipping loop (magenta) are highlighted, with other regions in grey. (B) Overlay of ribbon representations of the flipping loop and the motif II loop of chain B of apo and Asp-DACM-bound PvAspRS. The sidechain of residue R396 of apo PvAspRS is not fully built due to insufficient density. (C) B-factor analysis of the B chain flipping loops of apo PvAspRS and Asp-DACM-bound structures. The x-axis shows residue number. (D) 2F_o-F_c maps contoured at 2 σ (mesh surface) showing electron density supporting the position of Asp-DACM bound to chains A and B. (E) Ligplots of Asp-DACM active site interfaces for the A and B chains. Hydrogen bonds and salt bridges are depicted with dashed (green) lines. Other interactions between protein and ligand are indicated by red arcs.

https://doi.org/10.1371/journal.ppat.1013057.g005

The flipping loop, which plays an important role in substrate binding [39,40], is supported by a β-hairpin structure (Fig 5A, magenta). In apo PvAspRS, residues SEGG (351–354) of the Chain A flipping loop are not resolved (S7A Fig). In Chain B, the flipping loop can be built (Fig 5B). The SEGG loop interacts with a conserved residue Q371, as well as S350, N356 and A357 (S7B Fig). B-factor analysis suggests the SEGG loop is more mobile than neighboring regions of the protein (Fig 5C).

Structure of Asp-DACM-bound PvAspRS

We solved the structure of PvAspRS with bound synthetic Asp-DACM and refined it to 2.4 Å resolution (S3 Table). Well-defined electron density was observed for the adduct in both chains A and B (Fig 5D). The Asp-DACM is located in the active site pocket of the catalytic domain and interacts with a number of amino acid residues (Fig 5E). The adenine moiety is stacked between H405 and Y409 from motif II. H405 moves from its position in the apo protein to form a parallel π stack with the adenine ring (Figs 5B and S7A). Other motif II loop residues, R396 and E398, also adopt different conformations in the Asp-DACM structure. R396 is highly conserved in prokaryotic and eukaryotic enzymes; and is known to stabilize the transition state during tRNA charging [41]. In both chains, R396 extends to interact with the aspartate of Asp-DACM (Figs 5B,5E and S7A). Similarly, E398 repositions to form an H-bond with the adenine. The aspartic acid moiety of Asp-DACM also interacts with Q374, R561, and K377 (Fig 5E).

As for apo PvAspRS, the flipping loop in Chain B of Asp-DACM-bound PvAspRS is well defined. Residue E352 (of the SEGG motif in Chain B) repositions to interact with the amino group of the aspartate, leading to a small movement towards the ligand (Fig 5B). In Chain A of Asp-DACM-bound PvAspRS, two additional residues of the flipping loop (G353 and G354) are resolved compared with the apo protein (S7C Fig).

Structure of Asp-AMS-bound PvAspRS

We also solved the structure of PvAspRS with bound synthetic Asp-AMS and refined it to 1.8 Å resolution, where well-defined electron density was observed for the adduct (S3 Table and Fig 6A). Asp-AMS displays similar interactions to Asp-DACM (Fig 6B), consistent with their similar ability to thermally stabilize PvAspRS. Interestingly, in the Asp-AMS B chain structure, a Mg²⁺ ion coordinates the interaction between the sulfamate oxygen and E554 (Fig 6B). E352 interacts with the amino group of the aspartate in both chains, leading to stabilization of both flipping loops (Fig 6C). B-factor analysis reveals that the B chain flipping loop in Asp-AMS bound PvAspRS exhibits higher stability than in the apo structure (S7C Fig). Interestingly, H405 in the motif II loop adopts a different orientation to the Asp-DACM-bound structure (Fig 6D). It forms a less intimate contact with the adenine ring that may be mediated by CH-π or cation-π interactions. Notably, B-factor analysis indicates that, in both chains, the motif II loop in Asp-AMS-bound PvAspRS exhibits greater mobility than other structures (Fig 6E). Thus, the interactions of Asp-DACM and Asp-AMS with the active site are subtly different.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Crystal structure of Asp-AMS bound PvAspRS.

(A) 2F_o-F_c maps contoured at 2 σ (mesh surface) showing electron density supporting the position of Asp-AMS bound to chain A and chain B. (B) Ligplots of Asp-AMS/active site interfaces for the A and B chains. Hydrogen bonds and salt bridges are depicted with dashed (green) lines. Other interactions between protein and ligand are indicated by red arcs. (C) Overlay of chains A and B of Asp-AMS-bound PvAspRS showing the flipping loop and the motif II loop. (D) Overlay of chains A and B of Asp-AMS-bound and Asp-DACM-bound PvAspRS, showing the different positions of H405. (E) B-factor analysis of the motif II loops of the apo and ligand-bound structures. The x-axis shows residue number.

https://doi.org/10.1371/journal.ppat.1013057.g006

Structure of Asp-AMP-bound PvAspRS

We also solved the structure of PvAspRS with the natural intermediate, Asp-AMP, and refined it to 2.1 Å resolution (S3 Table and S8A Fig). Asp-AMP is located in the equivalent position within the catalytic domain as Asp-DACM and Asp-AMS; and many of the interactions are similar (S8B Fig). Of interest, in the Asp-AMP structure, H405 occupies the same position as in the apo protein and does not form a π-π interaction with the ligand (S8C Fig). Similarly, E398 from the motif II loop remains in a similar position as in the apo protein, directed away from the active site (S8C Fig, right panel). This orientation is not consistent with E398 interacting with the amino group in the adenine ring, as is the case for the Asp-DACM and Asp-AMS structures (compare Figs 6C and S8C, right panel). As for apo PvAspRS, the flipping loop SEGG residues are resolved in Chain B, but not in Chain A (S8C Fig). B-factor analysis shows that the B chain flipping loop in the Asp-AMP-bound structure has similar flexibility to the apo and Asp-DACM structures (S7C Fig); but is less ordered than the Asp-AMS structure (S7C Fig).

Comparison of apo PvAspRS with apo HsAspRS

The structure of apo HsAspRS (PDB: 4j15) has been published previously [42]. The two chains of the dimer are equivalent in the human structure. Superimposition of apo HsAspRS with apo PvAspRS (96–631) chain B reveals conservation of the overall structure, particularly in the catalytic and anticodon-binding domains (Fig 7A). By contrast, some regions of differential flexibility are observed (Fig 7B and 7C). Of particular interest, residues I224-Q248, comprising the flipping loop and flanking residues, are disordered in HsAspRS. By contrast, in the equivalent region of apo PvAspRS (L347-Q371), the flipping loop and flanking β-hairpin are both well-defined in Chain B (Fig 7B), while in apo Chain A, only the flipping loop itself is not resolved. Similarly, a loop (R273 to H282) in the middle of motif II, including residues that interact with the end of acceptor stem on the major groove [42–44], is not resolved in apo HsAspRS, but the equivalent region (R396-H405) is ordered in apo PvAspRS (Fig 7B). PvAspRS exhibits an insert in the anticodon-binding domain that is not present in HsAspRS. The insert is disordered and not resolved in the PvAspRS structure (Fig 7C).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Comparison of the crystal structures of apo PvAspRS with HsAspRS and EcAspRS.

(A) Overlay of apo PvAspRS and apo HsAspRS (PDB: 4j15) illustrating the conservation of the overall structure. (B) The flipping loop and associated β-hairpin region, and the motif II loop, are unstructured in apo HsAspRS and more ordered in apo PvAspRS (Chain B). (C) Overlay of the anticodon-binding domains of PvAspRS and HsAspRS. The Plasmodium-specific insertion is not resolved. (D) Overlay of the apo (top panel) and Asp-AMP-bound (bottom panel) PvAspRS (chain B) and EcAspRS (PDB: 1eqr chain A and 1il2 chain B) active sites illustrating the different positions of the flipping loops. (E) Comparison of the flipping loop interactions of Asp-AMP-bound PvAspRS (SEGG; chain B) and EcAspRS (PEGA; PDB: 1il2, chain B). SEGG interactions are depicted on the left panel. PEGA interactions are depicted in the right panel. SEGG and PEGA loop residues are shown as sticks and the interacting residues are shown in grey.

https://doi.org/10.1371/journal.ppat.1013057.g007

Comparison of apo and Asp-AMP-bound structures of PvAspRS and EcAspRS

We also compared the apo and Asp-AMP-bound structures of PvAspRS (96–631) with the equivalent E. coli AspRS structures. In apo EcAspRS (PDB: 1eqr; [45]), the flipping loop is resolved, but with a high B factor. The apo EcAspRS flipping loop (molecule 1/chain A) adopts an “open” conformation that is displaced compared with the flipping loop in chain B of PvAspRS (Fig 7D, top panel). By contrast, in the Asp-AMP-bound EcAspRS structure (PDB: 1il2; [39]), the flipping loop adopts a “closed” structure (Fig 7D, bottom panel) that is similar to that of Asp-AMP-bound PvAspRS. In apo PvAspRS, the SEGG loop is stabilized by interactions with the conserved residues, Q371 and S372, as well as residues Y532, S350 and A357 (Fig 7E). These latter residues are not conserved in EcAspRS, and the equivalent PEGA loop only interacts with Q192, S193 and two water molecules. Thus, differences in the sequences of PvAspRS and EcAspRS in this region likely underpin the observed structural differences, enabling the apo PvAspRS chain B flipping loop to adopt a conformation that is poised to form a “closed” structure.

Ethics statement

The use of human RBCs and serum in this study was ethically approved by the Office of Research Ethics and Integrity (The University of Melbourne). RBCs and serum were acquired from the Australian Red Cross Lifeblood service, Melbourne, Australia. Reference number: 2022-25738-35573-3.

Inhibition of the growth of P. falciparum cultures

Parasite-infected RBCs (3D7 strain [28]) were cultured in RPMI-HEPES containing 0.25% AlbuMAX II and 5% human serum, as described previously [55]. Sorbitol-synchronized ring stage parasites at 1% parasitaemia and 0.2% haematocrit (were incubated with DACM or AMS for 72 h and viability was assessed in the next cycle by flow cytometry, after labelling with 2 µM SYTO 61 (Thermo Fisher Scientific), as previously described [29]. The parasitemia was normalized to untreated and “kill-treated” controls, treated with 10x IC₅₀ concentration of each compound, for 72 h. Cells were pelleted at 400 g for 1.5 min and were incubated with 2 µM SYTO 61 in phosphate-buffered saline (PBS; Molecular Probes, Life Technologies) at RT for 15 min. Nine volumes of PBS were added to the cells (final SYTO 61 concentration: 0.2 µM) and incubated for a further 30 min at RT. The samples were analyzed by BD Biosciences FACSCanto II flow cytometer using the APC channel where the forward and side scatter was used to gate total cells. Data were processed by BD FACSDiva Software and FlowJo. The half maximal inhibitory concentration (IC₅₀) was determined using nonlinear regression (curve fit) in GraphPad Prism.

Cytotoxicity of PM03 against HepG2 cells

The HepG2 (Human Caucasian hepatocyte carcinoma) cell line was procured from Cell Repository NCCS, India, and cultured in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% Fetal Bovine Serum (FBS), 4 mM glutamine and 50 µg/mL penicillin-streptomycin, in a humidified incubator at 37°C with 5% CO_2. Using an assay procedure modified from [56], 10,000 cells per well were incubated in 96-well plates for 24 h at 37°C with 5% CO₂, to allow attachment. The medium was removed; and cells were treated with fresh medium (100 µL) containing either vehicle or serial dilutions (2-fold) of the compound, prepared in DMEM, with 2% FBS, in triplicate. Cells treated with 20% DMSO were used as a kill control. The cells were incubated for 48 h at 37°C with 5% CO₂. The growth medium was aspirated, and 100 µL of 0.5 mg/mL MTT solution was added to each well. After 3 h at 37°C, with 5% CO₂, the MTT solution was removed and 100% DMSO was added to dissolve formazan crystals, with shaking, in the dark, at 37°C for 15 min. The absorbance was measured at 570 nm using a SpectraMax M3 microplate reader. The absorbance at 630 nm was subtracted to correct for background noise. Graph Prism 9 was used to generate the dose-response curve using non-linear regression analysis (variable slope).

Protein translation assay

P. falciparum Cam3.II^Rev [57] trophozoite (30–35 h p.i.) infected RBCs (0.2% hematocrit and 1% parasitemia) were exposed to the relevant compounds for 6 h. The cells were labelled with O-propargyl-puromycin (OPP (Abcam); 4 µM) in the final 2 h, washed two times (PBS + 3% human serum) and then fixed and permeabilized, as described previously [17]. The click reaction was performed for 1 h at 37^oC in the presence of CuSO₄ (0.1 mM), tris-hydroxypropyltriazolylmethylamine (THPTA; 0.5 mM) and sodium ascorbate (5 mM) to bring about azide-alkyne cycloaddition to Alexa Fluor 488 azide (0.1 µM). Pellets were washed four times and resuspended in PBS + 3% human serum containing 25 μg/mL propidium iodide (Invitrogen). Flow cytometry (FACS Canto II; BD Biosciences, San Jose, CA) was performed using a BD FACSDiva (version 8.0) and FlowJo (version 10.9). Side scatter height (SSC-H) and forward scatter area (FSC-A) density plots were used to gate the total cell population. FSC-A and forward scatter width (FSC-W) plots were used to gate the single cell population. The FITC and PerCP-Cy5.5-H channels were used to detect the Alexa Fluor 488 and propidium iodide (PI) positive populations, representing parasitized RBCs. The same 6-h drug treatment conditions were set up in parallel for viability assessment. Cells were washed intensively 3 x with Complete Culture Medium (CCM) to remove inhibitors and returned to culture. Parasite viability was assessed in the next cycle as previously described above.

Phospho-eIF2α analysis

Trophozoite-infected RBCs (26 – 32 h p.i.; 2.5% hematocrit, 5–6% parasitemia) of the Cam3.II^Rev line [57] or an eIK1 knockout line [30] (kindly provided by Prof Christian Doerig, RMIT University), were incubated with 0.2 µM borrelidin, 1 µM ML901, 1 µM AMS, 1 µM DACM, or DMSO, for 3 h. The RBC pellet was washed 3 x in PBS + cOmplete EDTA-free Protease Inhibitor Cocktail. The pelleted cells were lysed by resuspension in PBS + 0.05% saponin, on ice. Washed pellets were resuspended in Bolt LDS sample buffer plus reducing agent and subjected to Western analysis as described previously [17]. Primary antibodies: rabbit anti-phospho-eIF2α (Cell Signaling Technology-119A11; 1:1,000); polyclonal mouse anti-PfBiP, generated using recombinant PfBiP at the WEHI Antibody Services (1:1,000). Secondary antibodies: goat anti-rabbit IgG-peroxidase (Chemicon-A132P; 1:20,000), goat anti-mouse IgG-peroxidase (Chemicon-A181P; 1:50,000). Membranes were washed three times in 1xPBS + 0.1% Tween 20. Chemiluminescence was detected using the Bio-Rad ChemiDoc MP imaging system.

Sample preparation to identify amino acid DACM conjugates

Late trophozoite stage P. falciparum (3D7 strain) culture samples was exposed to 10 μM DACM for 3 h. The parasite-infected RBCs were lysed with 0.1% saponin in PBS and the parasite pellet was washed 3 times with ice-cold PBS. Cell pellets were kept on ice and resuspended in water as one volume, followed by the addition of five volumes of cold chloroform-methanol (2:1 [vol/vol]) solution. Samples were incubated on ice for 5 min, subjected to vortex mixing for 1 min and centrifuged at 13,500 x g for 10 min at 4°C to form 2 phases. The top aqueous layer was transferred to a new tube and subjected to LC-MS analysis. The synthetic Asp-DACM standard was processed in the same way. Data analysis was performed using Xcalibur (version 4.4).

High-performance liquid chromatography (HPLC) and mass spectrometric (MS) analyses

Samples were analyzed by reversed-phase ultra-high performance liquid chromatography (UHPLC) coupled to tandem mass spectrometry (MS/MS) employing a Vanquish UHPLC linked to an Orbitrap Fusion Lumos mass spectrometer (Thermo Fisher Scientific, San Jose, CA, USA) operated in positive ion mode, modified from a previous procedure [15]. Solvent A was 0.1% formic acid acetate in water and solvent B was 0.1% formic acid in acetonitrile. 5 μL of each sample was injected onto an RRHD Eclipse Plus C18 column (2.1 × 100 mm, 1.8 μm; Agilent Technologies, USA) held at 50 °C with a solvent flow rate of 350 μL/min. The solvent gradient was as follows [Time (min), B %]: [0, 0], [6, 0], [13,25], [13.1, 99], [14, 99], [14.1, 0], [18, 0]. Mass Spectrometry experiments were performed using a Heated Electrospray Ionization (HESI) source. The spray voltage, flow rate of sheath, auxiliary and sweep gases were 3.5 kV, 25, 5, and 0 ‘arbitrary’ unit(s), respectively. The ion transfer tube and vaporizer temperatures were maintained at 300°C and 150°C, respectively, and the S-Lens RF level was set at 30%. A full-scan MS spectrum and targeted MS/MS for the proton adduct of Asp-DACM or the 20 possible common amino acid-containing inhibitor adducts were acquired in cycles throughout the run. The full-scan MS spectra were acquired in the Orbitrap at a mass resolving power of 120,000 (at m/z 200) across an m/z range of 200–1500 using quadrupole isolation and the targeted MS/MS were acquired using higher-energy collisional dissociation (HCD)-MS/MS in the Orbitrap at a mass resolving power of 7500 (at m/z 200), a stepped normalized collision energy (NCE) of 15, 30 and 45% and an m/z isolation window of 1.2.

Generation of recombinant aaRSs

The gene sequences encoding native length PfAspRS (49–626) (PlasmoDB ID: PF3D7_0102900), native length PvAspRS (51–631) (PVX_081610) and full-length HsAspRS (1–501) (NP_001340.2) were codon-optimized for expression in Escherichia coli, synthesized by GenScript, and cloned into pET-11a vector with a histidine tag and TEV cleavage site. Truncated PfAspRS (87–626) and PvAspRS (96–631) genes were amplified from the synthesized plasmids using PCR and cloned into pET-11a and pETM-41 vectors, respectively. The proteins were overexpressed in E. coli BL21 (DE3) using 0.05 mM or 0.1 mM IPTG induction at 16°C, overnight. The lysis buffer was 50 mM Tris pH 7.4 (or pH 8.0), 350 mM NaCl, 40 mM imidazole, 0.5 mM tris(2-carboxyethyl)phosphine (TCEP), 1 mg/mL lysozyme, and 1x complete protease inhibitor. Clarified lysate was loaded onto a HisTrap (HP) 5 mL nickel column (Cytiva). Proteins were eluted with a gradient of Buffer B (50 mM Tris pH 7.4, 350 mM NaCl, 500 mM imidazole, 0.5 mM TCEP). PfAspRS was eluted at ~40% Buffer B. The eluted sample was dialyzed against 50 mM Tris-HCl pH 7.4, 150 mM NaCl, 25 mM imidazole (or 200 mM imidazole), and 1 mM TCEP, with TEV protease to remove the N- terminal histidine tag. The sample was concentrated and subjected to size exclusion chromatography using buffer containing 25 mM or 50 mM Tris-HCl pH 8.0, 150 mM NaCl, and 0.5 or 1 mM TCEP.

Characterization of native length PfAspRS

For analytical ultracentrifugation, recombinant PfAspRS was prepared at 1, 0.6, and 0.2 mg/mL in buffer containing 50 mM Tris pH 8.0, 150 mM NaCl, 1 mM TCEP. Samples and buffer (reference solution) were centrifuged at 200,000 g at 20°C, in a Beckman Coulter XL-I analytical ultracentrifuge, equipped with UV-visible scanning optics. Radial absorbance data were monitored and collected at a wavelength of 290 nm. Sedimentation data were fitted to a continuous sedimentation coefficient (c(s)) model, with frictional ratios estimated using SEDFIT software [58].

For mass spectrometry, purified protein (5 μg) in 25 mM Tris pH 8.0, 150 mM NaCl, 1 mM TCEP was subjected to an Agilent 1200 HPLC, equipped with a C18 column, connected to an Agilent 6220 Accurate-Mass electrospray ionization time-of-flight (ESI-TOF) mass spectrometer. Data acquisition and analysis were performed using Mass Hunter Software (Agilent).

Preparation of E. coli tRNA

Total tRNA from E. coli was isolated with modifications from a previous report [59]. E. coli BL21(DE3) cells were cultured in 2 × Yeast Extract Tryptone medium at 37°C overnight. Cells were harvested by centrifugation, resuspended in diethyl pyrocarbonate (DEPC)-treated water, and lysed by adding TRIzol reagent (Invitrogen) at a 3:1 ratio to the cell suspension, followed by vigorous vortexing. After centrifugation, the aqueous phase was extracted, and acid-phenol: chloroform pH 4.5 (with indole-3-acetic acid (IAA), 25:24:21; Invitrogen) was added. Following another centrifugation step, the supernatant was collected. This step was repeated until a clear interface was observed. Next, LiCl was added to a final concentration of 1 M. tRNA was precipitated with ice-cold isopropanol, then dissolved in DEPC-treated water for further use.

ATP consumption assay

The consumption of ATP by native length PfAspRS, native length PvAspRS, truncated PvAspRS and full-length HsAspRS was determined using a luciferase-based assay as per the manufacturer’s instructions (Kinase-Glo Luminescent Kinase Assay, Promega). Reactions were conducted in 25 mM Tris-HCl (pH 8), 150 mM NaCl, 5 mM MgCl₂, 0.1 mg/mL BSA, 1 mM TCEP, with 200 µM L-aspartate, 10 µM ATP, 1 unit/mL inorganic pyrophosphatase and 80 or 160 μM E.coli tRNA (if present). Enzyme concentration and incubation time for each experiment are described in the relevant figure legend. Reactions were incubated at 37°C, followed by addition of the Kinase-Glo reagent. Luminescence output was measured using a plate reader (CLARIOstar, BMG LABTECH) and the highest signal within 20 min after addition of reagents was recorded using MARS data analysis software (version 3.32). Assay conditions were optimized to ensure ATP consumption is in the linear range with respect to AspRS concentration. The concentration of ATP was quantified by linear regression using an ATP standard curve (Microsoft Excel).

Differential scanning fluorimetry (DSF)

The effect of DACM, AMS and Asp-DACM on the thermal stability of AspRS enzymes was assayed as previously described [15]. Briefly, the relevant AspRS (1.5 μM) was incubated in the presence or absence of 10–50 μM DACM or AMS, or 2, 5 or 10 μM Asp-DACM, with 10 μM ATP, 20 μM L-aspartate, 80 μM EctRNA, in 25 mM Tris-HCl (pH 8), 150 mM NaCl, 5 mM MgCl₂, 1 mM TCEP, at 37°C for 3 h. SYPRO Orange (Sigma-Aldrich; 5,000X concentrate in DMSO) was added to the reaction mixture at a final concentration of 5X. 25 μL of the sample was added into each well of a 96-well qPCR plate (Applied Biosystems). The plate was sealed and analyzed using StepOnePlus Real-Time PCR system (Applied Biosystems). The samples were heated from 20°C to 90°C with a 1°C per min continuous gradient. The thermal unfolding curve was plotted as the first derivative curve of the raw fluorescence values. The melting temperature (T_m), defined as the peak of the first derivative curve, was used to assess the thermal stability of protein-ligand complexes.

Crystallization and X-ray diffraction data collection

For crystallization, PvAspRS (96–631) in the apo form, prepared in 20 mM HEPES pH 8.0, 200 mM NaCl, 10% glycerol, and 5 mM beta-mercaptoethanol, was concentrated to 20 mg/mL and incubated without or with bound natural substrates (ATP and L-Asp) or synthetic Asp-AMS or Asp-DACM, with 5 mM MgCl₂, at a molar ratio of 1:4–1:30 (PvAspRS monomer/ ligand). Crystallization experiments were performed using the sitting-drop or hanging-drop vapour-diffusion method at 293 K. Initial crystallization screening was carried out in a 96-well plate (Corning, Lowell, Massachusetts, USA) with ViewDrop II seals (SPT LabTech, Melbourn, England) using the commercially available crystallization sparse-matrix screens (SG1, ProPlex, and Morpheus I and II; Molecular Dimensions, UK) [60]. Three different drop ratios were aliquoted using a Mosquito nanolitre dispenser system (TTP LabTech, Melbourn, England) or an NT8 drop setter (Formulatrix) by mixing protein and reservoir solutions at 1:1, 2:1 and 1:2 drop ratios; final volume 150 nL). The crystallization droplets were equilibrated against a 75 µL reservoir solution. Initial crystals of PvAspRS were obtained in the following conditions. Apo PvAspRS: SG1-D10 (0.2 M lithium sulfate, 0.1 M Bis-Tris pH 6.5 and 25% w/v PEG 3350); PvAspRS (Asp-AMP): SG1-C12 (0.2 M sodium acetate trihydrate, 0.1 M Bis-Tris pH 5.5 and 25% w/v PEG 3350); PvAspRS (Asp-AMS): Morpheus I-A9 (0.06 M divalents (0.3 M magnesium chloride hexahydrate; 0.3 M calcium chloride dihydrate), 0.1 M Buffer System (Tris base and BICINE), pH 8.5 and 30% v/v Precipitant Mix (40% v/v PEG 500 MME and 20% w/v PEG 20000); PvAspRS (Asp-DACM): Morpheus II-B5 (15% (w/v) PEG 3000, 20% (v/v) 1,2,4-butanetriol, 1% (w/v) nondetergent sulfobetaine (NDSB) 256, 0.5 mM manganese chloride, 0.5 mM cobalt chloride, and 0.5 mM zinc chloride). The crystals were cryoprotected using 10–20% glycerol before being flash-cooled in liquid nitrogen. X-ray diffraction experiments were conducted on the MX2 beamline at the Australian Synchrotron [61] or the I03 beamline at the Diamond Light Source, UK (S3 Table).

Structure determination

Several datasets were collected. Data were indexed and integrated using the XDS software package [62] and scaled using AIMLESS [63]. Alternatively, data were processed using the xia2/DIALS [64] and autoPROC [65] pipeline. Human AspRS (HsAspRS, PDB ID: 4J15 [42]) was used as the phasing model. The initial phases were determined by molecular replacement using PHASER [66] or Auto-Rickshaw [67]. The model was further refined using phenix.refine from PHENIX [68,69] and manually built using COOT [70]. Ligands, ions and water molecules were added to their electron densities after several rounds of manual model building and refinement. Structure refinement was performed using non-crystallographic torsion restraints and translation/libration/screw (TLS) refinement with each chain comprising a single TLS group. Restraints for Asp-AMP, Asp-AMS, Asp-DACM were generated using phenix.elbow [71] or GRADE Web Server (Global Phasing Ltd, https://grade.globalphasing.org). Difference density peaks observed near the 2-Cl moiety of Asp-DACM suggested radiation damage in this location during data collection. MolProbity [72,73] in PHENIX suite was used to evaluate model quality and figures were generated using UCSF Chimera [74] and PyMOL (http://www.pymol.org), including the embedded PyMOL secondary structure assignment. Magnesium ions were identified using CheckMyMetal [75]. Complete data collection and refinement statistics are summarized in S3 Table. LigPlot+ (version 2.2.8) was employed to analyze ligand-protein interactions and to generate 2D graphical maps [76].

Isotropic B-factor analysis

Isotropic B-factors for the alpha carbons for each residue were extracted from the PDB files in PyMOL Version 2.5.4 [77]. The B-factors were corrected by dividing by the Wilson B-factor of each structure and then normalised using the following equation: . The resulting normalized, dimensionless B-factor derived values ranged from 0-1, with higher values indicating greater atomic mobility.

Chemistry

Synthetic procedures and compound characterizations are provided in S1 Text.